Iterated Fact Sheets With R Markdown

Aaron R. Williams (@awunderground)

Urban Institute

Who we are

• 300 researchers
• 50 years

What we do

Our mission is to open minds, shape decisions, and offer solutions through economic and social policy research.

Our objective research helps expand opportunities for all, reduce hardship among the most vulnerable, and strengthen the effectiveness of the public sector.

• Tax Policy Center analysis of candidate tax plans
• Health Policy Center analysis of insurance coverage (King vs. Burwell)
• Pioneers of paired testing and microsimulation

Motivation

Getting information into the right hands is important.

Motivation

Our reports often seem like this.

Motivation

Our attention spans are trained for this.

## [1] "https://imgs.xkcd.com/comics/boyfriend.png"

Economic and social policy comic strips

Economic and social policy comic strips Iterated fact sheets and web pages

What does this look like?

1. Targeted and brief

What does this look like?

2. Visually appealing (and branded)

Urban Institute data visualization style guide

What does this look like?

3. Not labor intensive

Examples

• State fact sheets
• County fact sheets
• State fact pages

R Markdown

.Rmd

• Text and images
• Code
• Code output

.Rmd

• Text and images
• Code
• Code output

R4DS

Three ingredients for fact sheets/pages

1. Data structure
2. Parameterized .Rmd template
3. Iterating script

1) Data structure

• Typically one row per fact sheet/page

2) Parameterized .Rmd template

Parameters for R Markdown documents are specified in the YAML header:

---
params:
  state: "Alabama"
  year: 2016
---

params$state can be used anywhere in the document as "Alabama".

First, use the parameters to filter the big data structure to just data for an individual state or locality.

data <- data %>%
  filter(state == params$state)

2) Parameterized .Rmd template

---
params:
  state: "Alabama"
  year: 2016
---

Now, we can use data$<variable> for the state/locality specific variables and params$<parameter> throughout the document:

There were 'r data$fatalities' traffic fatalities in 'r params$state' in 'r params$year'. 

There were 1,038 traffic fatalities in Alabama in 2016.

2) Parameterized .Rmd template

Data viz!

2) Parameterized .Rmd template

• LaTeX for .pdf

• preamble.tex
• general document styles
• LaTeX macros
  • urbntitle{}
  • urbnboilerplate{}

2) Parameterized .Rmd template

• CSS for .html
• CSS for .md

• styles.css

3) Iterating script

# create a data frame with parameters and output file names
runs <- tibble(
  # creates a string with output file names in the form <index>.pdf
  output_file = str_c(index, ".html"),    
  # creates a nested list of parameters for each object in the index
  params = map(index, ~list(state = .))  
)

# iterate render() along the tibble of parameters and file names
runs %>%
  pwalk(rmarkdown::render, input = "template.Rmd", output_dir = "factsheets")

3) Iterating script

TIP! Don’t use NA in a YAML header.

---
params:
  state: "Alabama"
  year: 2016
---

Fact Sheet Tool

Fact Page Tool

Overall Process

1. Why fact sheets?
2. Create a Word/Google Doc template (if collaborating)
3. Recreate the template for one iteration in .Rmd
4. Iterate the .Rmd
5. Quality control
6. Create a home for the fact sheets

Takeaways

• Visually consistent
• Accurate
• Easy to update

Data@Urban

• Iterated fact sheets with R Markdown
• Iterated PDFs with R Markdown

Reach out to me

• awilliams@urban.org
• @awunderground

References

NHTSA Traffic Safety Fact Sheets

Texas Statehouse - LoneStarMike

Encyclopedia - WestportWiki